CN-121998166-A - Method for predicting wind speed of bridge site in canyon area based on valley wind theory and terrain feature decomposition technology

CN121998166ACN 121998166 ACN121998166 ACN 121998166ACN-121998166-A

Abstract

The invention provides a method for predicting the wind speed of a canyon area bridge site based on a valley wind theory and a topography feature decomposition technology, which comprises the steps of obtaining observed wind speed and wind direction data and meteorological element data, obtaining high-resolution topography elevation data nearby the bridge site, calculating topography feature parameters, determining the canyon trend of the position of an observation site according to coordinates and topography elevation data of the observation site, decomposing an observed wind speed and wind direction time sequence according to the canyon trend to respectively obtain axial wind along the canyon trend and normal wind perpendicular to the canyon trend, predicting the axial wind component and the normal wind component of a target site by utilizing an axial wind prediction model and a normal wind prediction model, and carrying out projection synthesis on the prediction result along the canyon trend to obtain a final prediction result of wind speed and wind direction.

Inventors

LI FEI
LI PENG
Zhu jinhuan
ZHOU LIBO
HE XUHUI

Assignees

中国科学院大气物理研究所
中南大学
中国国家铁路集团有限公司

Dates

Publication Date: 20260508
Application Date: 20251226

Claims (9)

1. The method for predicting the wind speed of the bridge site in the canyon area based on the valley wind theory and the terrain feature decomposition technology is characterized by comprising the following steps: The method comprises the steps of obtaining observed wind speed and wind direction data of stations nearby a bridge site and meteorological element data of standard meteorological stations around the bridge site, wherein the meteorological element data at least comprise temperature, humidity, air pressure, wind speed and wind direction, total radiation, long-wave radiation flux, short-wave radiation flux, sensible heat flux and latent heat flux; The method comprises the steps of obtaining high-resolution topographic elevation data near a bridge site, and calculating topographic feature parameters, wherein the topographic feature parameters comprise coordinates of an observation site, a topographic gradient in a canyon area, a mountain slope inclination angle, a canyon depth and a canyon width; Respectively projecting wind speed and wind direction data obtained from an observation station along the axial direction of the canyon and the normal direction of the canyon to obtain an axial wind component and a normal wind component; Correcting based on sunrise time and sunset time of the site of the bridge site by combining with the topographical features of the canyon, acquiring switching time of mountain wind and valley wind, and constructing a mountain valley wind switching time sequence; The method comprises the steps of taking an axial wind component as a target variable, taking the mountain valley wind switching time sequence and the standard weather station weather element time sequence as characteristic factors, constructing an axial wind prediction model based on an LSTM model, taking a normal wind component as a target variable, taking the mountain valley wind switching time sequence and the standard weather station weather element time sequence as characteristic factors, and constructing the normal wind prediction model based on the LSTM model; The axial wind prediction model and the normal wind prediction model are utilized to respectively predict the axial wind component and the normal wind component of a target site, and projection synthesis is carried out on the prediction results along the canyon trend to obtain the final prediction results of wind speed and wind direction; And comparing the predicted wind speed result with the actual observed value, and quantitatively evaluating the prediction precision through an error index.
2. The method of claim 1, wherein determining the canyon trend and the canyon normal at the location of the observation site based on the coordinates of the observation site and the terrain elevation data comprises: Cutting off the terrain elevation data (namely cutting off the terrain DEM data with the side length of 15 km) by taking the bridge site center point as the center; extracting a canyon region, calculating a terrain gradient, and extracting a lowest value of a terrain height connecting line in the canyon region as a canyon bottom line; The canyon trend is calculated by linear regression as the canyon axis and the direction perpendicular to the canyon axis as the canyon normal.
3. The method according to claim 1, wherein the method of calculating the axial wind component and the normal wind component is: clockwise rotation with 0 degree of positive north direction is positive, positive north direction is positive y-axis direction, positive east direction is positive x-axis direction, wind speed is V, and meteorological wind direction angle is theta, then wind direction and positive x-axis direction included angle The method comprises the following steps: ; Based on the orthogonal method decomposition: ; ; If the angle between the axial direction (trend) of the canyon and the positive direction of the x-axis is Then the component in the axial direction ; Component along the normal (perpendicular to the axial) of the canyon 。
4. The method according to claim 1, wherein the correcting is performed based on the sunrise time and the sunset time of the site of the bridge site in combination with the canyon topography feature, the switching time of the mountain wind and the valley wind is obtained, and the mountain wind switching time series is constructed, comprising the correcting steps of: calculating sunrise time and sunset time of the site of the bridge site based on the longitude and latitude as references; Correcting sunrise time and sunset time according to the canyon width and canyon depth, the slope gradient, the slope direction and the altitude of the topography at two sides, obtaining mountain wind and valley wind switching time, and constructing a mountain valley wind switching time sequence.
5. The method of claim 4, wherein calculating the sunrise time and the sunset time of the site of the bridge based on the longitude and latitude as the reference, and correcting the sunrise time and the sunset time according to the width of the canyon and the depth of the canyon, the slope of the terrain on both sides, the slope direction and the altitude to obtain the switching time of the mountain wind and the valley wind, and constructing the switching time sequence of the mountain wind comprises the following steps: calculating daily sunrise time and sunset time, wherein the initial correction time is 0; correcting according to the canyon width and the canyon depth, and if the canyon width is smaller than 800 meters or the canyon depth is larger than 500 meters, delaying the correction time of the valley wind generation time by 60 minutes; Correcting according to the terrain gradient, and if the terrain gradient is larger than 30 degrees, advancing the correction time of the valley wind generation time by 30 minutes; Correcting according to the slope direction, if the slope direction is north, the slope direction is positive, and correcting the valley wind generation time for 60 minutes later; correcting according to altitude, correcting and delaying the generation time of mountain valley wind for 30 min every 1000m rise, coding and marking the state of mountain valley wind by cosine and sine functions, and converting sunrise and sunset time and correction time into LSTM model identifiable characteristics.
6. The method according to claim 1, wherein the lattice search and the optimal feature selection method are used to select optimal model parameters and optimal feature combinations when constructing the axial wind prediction model and the normal wind prediction model.
7. The method of claim 6, wherein in constructing the axial wind prediction model and constructing the normal wind prediction model, further comprising: Taking site axial wind to be predicted as a target variable, taking a valley wind switching time sequence and a standard meteorological station element time sequence as characteristic factors, dividing a training set and a testing set, selecting an optimal model parameter and an optimal characteristic combination through a lattice point searching and optimal characteristic selecting method, enabling root mean square error of an actual value and a predicted value to be minimum, finally, until the model is converged, and constructing an axial wind prediction model based on an LSTM model; The site normal wind to be predicted is used as a target variable, the valley wind switching time sequence and the standard meteorological station element time sequence are used as characteristic factors, a training set and a testing set are divided, an optimal model parameter and an optimal characteristic combination are selected through a lattice point searching and optimal characteristic selecting method, so that the root mean square error of an actual value and a predicted value is minimum, finally, until the model converges, a normal wind prediction model is built based on an LSTM model.
8. The method of claim 1, wherein the projection synthesis method is to calculate the wind speed and wind direction angle by vector synthesis based on the predicted axial wind component and normal wind component, and the canyon strike angle.
9. The method of claim 1, wherein the error indicator comprises at least one of a root mean square error, an absolute deviation, and a relative error percentage.

Description

Method for predicting wind speed of bridge site in canyon area based on valley wind theory and terrain feature decomposition technology Technical Field The invention relates to the technical field of disaster monitoring and early warning, in particular to a method for predicting the wind speed of a bridge site in a canyon area based on a valley wind theory and a terrain feature decomposition technology. Background The area covered by the high-speed railway is gradually expanded from the past area mainly in the middle east plain to the area of the middle west complex mountain. The mountainous terrain is variable, and high-speed railways often need to traverse railroad bridges above canyons when passing through the mountainous canyons. The complicated mountain area has steep topography, the surface state is complex and various, and a strong and complicated regional and local scale circulation system can be formed under the strong solar radiation drive and the forced action of the topography. The strong and special local wind field in the complex mountain area can have great influence on the design, construction, operation and maintenance of the railway bridge and can form a great threat to the driving safety on the bridge. In order to ensure the running safety of the high-speed railway when passing through the mountain railway bridge, accurate prediction of wind near the bridge site is required. At present, accurate ultra-short-term prediction of wind speed sequences is still a difficult problem, mainly because wind is influenced by a plurality of factors such as temperature, air pressure and topography, and has complex change characteristics such as intermittence, volatility and randomness. Under complex terrain conditions, weather background wind fields, local atmospheric circulation, terrain dynamic forces, surface friction and thermal effects all contribute to wind formation. Among the commonly observed wind speed values, there are contributions from both the weather background wind field and the local atmospheric flows. The mountain area circulation structure is complex, and the mountain area circulation can comprise different systems. When only the thermodynamic effect of mountain terrain is considered, mountain wind systems consisting of slope wind, valley wind and mountain-plain wind exist in large-scale mountains, and the dominant areas are different. Slope wind is inclined and developed along the slope, the slope wind circulation time and space scale are minimum, the radiation change is responded immediately, the radiation heating is gradually transmitted to the inside of a valley, the temperature heating (cooling) speed is higher than that of a plain area in the valley due to small environment capacity, after a period of time, upward valley wind (downward valley wind from Gu Naliu to the plain) flowing from the plain to the valley along the valley axis occurs, mountain land-plain wind circulation is generated by the thermal difference between the whole mountain and the plain, and the circulation is difficult to distinguish due to confusion of the circulation in practical observation. The local atmospheric circulation movement of the complex terrain is mainly represented by valley wind, and the prominent characteristic of the valley wind is that the daily variation of wind speed, the valley wind during the day and the mountain wind at night are alternately generated. The formation mechanism of the valley wind is similar to that of the slope wind, namely, the distance between the air of the valley and the ground is short, the heating (cooling) rate is high, the air in the valley with the same height is far from the ground, the heating (cooling) rate is low, the isotherm is inclined along the slope, the buoyancy item enables the air flow to rise (sink) along the slope, and the valley wind is a three-dimensional structure. In the day-night alternation process, air temperature difference is generated between hills, valleys and mountain lands, the air temperature difference brings about density and air pressure difference of the ground atmosphere, and air pressure gradient force pushes air flow to move from a high-pressure (low-temperature) area to a low-pressure (high-temperature) area. Mountain wind and valley wind are caused by different heat effects in mountain top and valley day by day in mountain area. During daytime, in areas with higher elevations at the bottoms of valleys, the temperature is higher due to the irradiation of the sun, the air expands and the air pressure is reduced, so that the air climbs from low to high along the valley lines in the valleys to generate valley wind. The temperature is higher from the valley to the hillside in the afternoon, the wind force is stronger, the wind force reaches 3 levels when the afternoon wind force is maximum, the wind force is gradually weakened later, the hillside area is higher due to the topography and faster in heat dissipation at night, the ai